---

Aims: This research aimed to evaluate the spatial patterns of water quality and its controlling factors in the Mazandaran coastal ecosystem during winter using the multivariate analysis methods.
Materials and methods: Water quality parameters such as nutrients, temperature, conductivity, salinity, DO, pH, chlorophyll-a and turbidity were measured monthly in 16 stations (44 layers) along 4 transects (Amirabad, Babolsar, Noushahr and Ramsar). To evaluate the data, several multivariate statistical methods were used including discriminant function analysis, cluster and factor analysis as well as correlation test.
Findings: Results of cluster analysis showed that the sampling sites (44 layers) were classified into 4 groups. Based on discriminant analysis, 93.20% of the sampling sites correctly classified. Factor analysis extracted 4 principal components that explained 74.05% of the total variance. Based on these analyses, organic phosphorus, organic nitrogen, turbidity, chlorophyll-a and temperature were the most effective parameters on the spatial variation of water quality.
Conclusion: This study suggested that the number of sampling locations could be reduced to 3 transects including Amirabad, Babolsar and west coasts (Noushahr and Ramsar) and 2 stations (one surface layer and one deep layer). Transport of nutrients from land, sea floor and fish cage culture were the most effective factors on spatial patterns of water quality in Mazandaran coasts. Based on the results, multivariate statistical methods are also introduced as one of the useful methods for identifying the spatial pattern of water quality.